Flexure-based grippers offer an attractive alternative to conventional grippers used in robotics and automation. However, most existing designs appear to suffer from insufficient range of motion, loadability and support stiffness. This paper presents an approach to obtain well-performing flexure hinges for compact anthropomorphic grippers made via metal additive manufacturing. We propose a flexure hinge architecture that achieves a high range of motion despite the challenging combination of a small design space, high Young's modulus and limited minimum feature size. Furthermore, we present an optimization procedure to generate suitable tendon-driven designs with high loadability. Using this framework, a flexure hinge with an outer diameter of 21.5 mm and range of motion of ±30 deg is synthesized. For the range of 0 to 30 deg simulations show a lateral loadability of 52.5 to 18.6 N and lateral support stiffness of 12309 to 11130 N/m, determined at a gripper interface located 41.2 mm from the hinge pivot axis. Experiments confirm a loadability of at least 15.4 N and determined a stiffness of 8982 to 9727 N/m for same conditions. The results show that the flexure hinge architecture has large potential for a wide range of applications, while in combination with the optimization procedure superior designs for tendon-driven grippers can be obtained.
{"title":"Flexure hinge design and optimization for compact anthropomorphic grippers made via metal additive manufacturing","authors":"M. Tschiersky, Jan De Jong, Dannis Brouwer","doi":"10.1115/1.4063362","DOIUrl":"https://doi.org/10.1115/1.4063362","url":null,"abstract":"\u0000 Flexure-based grippers offer an attractive alternative to conventional grippers used in robotics and automation. However, most existing designs appear to suffer from insufficient range of motion, loadability and support stiffness. This paper presents an approach to obtain well-performing flexure hinges for compact anthropomorphic grippers made via metal additive manufacturing. We propose a flexure hinge architecture that achieves a high range of motion despite the challenging combination of a small design space, high Young's modulus and limited minimum feature size. Furthermore, we present an optimization procedure to generate suitable tendon-driven designs with high loadability. Using this framework, a flexure hinge with an outer diameter of 21.5 mm and range of motion of ±30 deg is synthesized. For the range of 0 to 30 deg simulations show a lateral loadability of 52.5 to 18.6 N and lateral support stiffness of 12309 to 11130 N/m, determined at a gripper interface located 41.2 mm from the hinge pivot axis. Experiments confirm a loadability of at least 15.4 N and determined a stiffness of 8982 to 9727 N/m for same conditions. The results show that the flexure hinge architecture has large potential for a wide range of applications, while in combination with the optimization procedure superior designs for tendon-driven grippers can be obtained.","PeriodicalId":50137,"journal":{"name":"Journal of Mechanical Design","volume":"28 1","pages":""},"PeriodicalIF":3.3,"publicationDate":"2023-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"75415835","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Many design problems involve reasoning about points in high-dimensional space. A common strategy is to first embed these high-dimensional points into a low-dimensional latent space. We propose that a good embedding should be isometric---i.e., preserving the geodesic distance between points on the data manifold in the latent space. However, enforcing isometry is non-trivial for common Neural embedding models such as autoencoders. Moreover, while theoretically appealing, it is unclear to what extent is enforcing isometry necessary for a given design analysis. This paper answers these questions by constructing an isometric embedding via an isometric autoencoder, which we employ to analyze an inverse airfoil design problem. Specifically, the paper describes how to train an isometric autoencoder and demonstrates its usefulness compared to non-isometric autoencoders on the UIUC airfoil dataset. Our ablation study illustrates that enforcing isometry is necessary for accurately discovering clusters through the latent space. We also show how isometric autoencoders can uncover pathologies in typical gradient-based Shape Optimization solvers through an analysis on the SU2-optimized airfoil dataset, wherein we find an over-reliance of the gradient solver on angle of attack. Overall, this paper motivates the use of isometry constraints in Neural embedding models, particularly in cases where researchers or designers intend to use distance-based analysis measures to analyze designs within the latent space. While this work focuses on airfoil design as an illustrative example, it applies to any domain where analyzing isometric design or data embeddings would be useful.
{"title":"Characterizing Designs via Isometric Embeddings: Applications to Airfoil Inverse Design","authors":"Qiuyi Chen, M. Fuge","doi":"10.1115/1.4063363","DOIUrl":"https://doi.org/10.1115/1.4063363","url":null,"abstract":"\u0000 Many design problems involve reasoning about points in high-dimensional space. A common strategy is to first embed these high-dimensional points into a low-dimensional latent space. We propose that a good embedding should be isometric---i.e., preserving the geodesic distance between points on the data manifold in the latent space. However, enforcing isometry is non-trivial for common Neural embedding models such as autoencoders. Moreover, while theoretically appealing, it is unclear to what extent is enforcing isometry necessary for a given design analysis. This paper answers these questions by constructing an isometric embedding via an isometric autoencoder, which we employ to analyze an inverse airfoil design problem. Specifically, the paper describes how to train an isometric autoencoder and demonstrates its usefulness compared to non-isometric autoencoders on the UIUC airfoil dataset. Our ablation study illustrates that enforcing isometry is necessary for accurately discovering clusters through the latent space. We also show how isometric autoencoders can uncover pathologies in typical gradient-based Shape Optimization solvers through an analysis on the SU2-optimized airfoil dataset, wherein we find an over-reliance of the gradient solver on angle of attack. Overall, this paper motivates the use of isometry constraints in Neural embedding models, particularly in cases where researchers or designers intend to use distance-based analysis measures to analyze designs within the latent space. While this work focuses on airfoil design as an illustrative example, it applies to any domain where analyzing isometric design or data embeddings would be useful.","PeriodicalId":50137,"journal":{"name":"Journal of Mechanical Design","volume":"31 1","pages":""},"PeriodicalIF":3.3,"publicationDate":"2023-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"73127229","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This paper presents a 3D material mask overlay topology optimization approach using truncated octahedron elements and spheroidal masks. Truncated octahedron elements provide face connectivity between two juxtaposed elements, thus, eliminating singular solutions inherently. A novel meshing scheme with Tetra-Kai-Decaheral or TKD (generic case of truncated octahedron) elements is proposed. The scheme is extended to parameterized generic-shape domains. Various benefits of implementing the elements are also highlighted, and the corresponding finite element is introduced. Spheroidal negative masks are employed to determine the material within the elements. Seven design variables define each mask. A material density formulation is proposed, and sensitivity analysis for gradient-based optimization is developed. fmincon MATLAB function is used for the optimization. The efficacy and success of the approach are demonstrated by solving structures and compliant mechanism design problems. Compliance is minimized for the former, whereas a multi-criteria arising due to flexibility and stiffness measures is extremized for optimizing the mechanisms. Convergence of the optimization is smooth. The volume constraint is satisfied and remains active at the end of the optimization.
{"title":"3D material mask overlay topology optimization approach with truncated-octahedron elements","authors":"Nikhil Singh, Prabhat Kumar, A. Saxena","doi":"10.1115/1.4063361","DOIUrl":"https://doi.org/10.1115/1.4063361","url":null,"abstract":"\u0000 This paper presents a 3D material mask overlay topology optimization approach using truncated octahedron elements and spheroidal masks. Truncated octahedron elements provide face connectivity between two juxtaposed elements, thus, eliminating singular solutions inherently. A novel meshing scheme with Tetra-Kai-Decaheral or TKD (generic case of truncated octahedron) elements is proposed. The scheme is extended to parameterized generic-shape domains. Various benefits of implementing the elements are also highlighted, and the corresponding finite element is introduced. Spheroidal negative masks are employed to determine the material within the elements. Seven design variables define each mask. A material density formulation is proposed, and sensitivity analysis for gradient-based optimization is developed. fmincon MATLAB function is used for the optimization. The efficacy and success of the approach are demonstrated by solving structures and compliant mechanism design problems. Compliance is minimized for the former, whereas a multi-criteria arising due to flexibility and stiffness measures is extremized for optimizing the mechanisms. Convergence of the optimization is smooth. The volume constraint is satisfied and remains active at the end of the optimization.","PeriodicalId":50137,"journal":{"name":"Journal of Mechanical Design","volume":"45 1","pages":""},"PeriodicalIF":3.3,"publicationDate":"2023-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"82134090","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
As computer-aided design (CAD) tools have become an essential aspect of modern mechanical engineering design, the demand for CAD experts has increased significantly. The development from novice, to proficient, to expert user is of particular interest to the industrial and academic design communities. Yet little is known about the development or characteristics of expert CAD skill; much of the past work that reports user action data is based on student or novice data. We compared the CAD modelling process across nine proficient and ten expert designers as they were tested to complete the same design task. Under identical conditions – the same time constraints in the same CAD platform and with the same task -- the expert users were able to complete a larger proportion of the task with higher dimensional accuracy. While the experts were able to dissect and retrieve geometries from manufacturing drawings more efficiently than proficient users, they were also able to plan a modelling strategy that required less effort and revisions. With our experimental findings, we identify the demand for procedural knowledge-building for young engineers, with the ultimate goal of more effectively developing experts in engineering design with CAD.
{"title":"What sets proficient and expert users apart? Results of a Computer-Aided Design experiment","authors":"Yuan Deng, James Chen, A. Olechowski","doi":"10.1115/1.4063360","DOIUrl":"https://doi.org/10.1115/1.4063360","url":null,"abstract":"\u0000 As computer-aided design (CAD) tools have become an essential aspect of modern mechanical engineering design, the demand for CAD experts has increased significantly. The development from novice, to proficient, to expert user is of particular interest to the industrial and academic design communities. Yet little is known about the development or characteristics of expert CAD skill; much of the past work that reports user action data is based on student or novice data. We compared the CAD modelling process across nine proficient and ten expert designers as they were tested to complete the same design task. Under identical conditions – the same time constraints in the same CAD platform and with the same task -- the expert users were able to complete a larger proportion of the task with higher dimensional accuracy. While the experts were able to dissect and retrieve geometries from manufacturing drawings more efficiently than proficient users, they were also able to plan a modelling strategy that required less effort and revisions. With our experimental findings, we identify the demand for procedural knowledge-building for young engineers, with the ultimate goal of more effectively developing experts in engineering design with CAD.","PeriodicalId":50137,"journal":{"name":"Journal of Mechanical Design","volume":"30 1","pages":""},"PeriodicalIF":3.3,"publicationDate":"2023-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"74308101","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Delta-like architectures are widely used for fast pick-and-place applications. When rotational degrees of freedom are required to perform a task, one or more UPU kinematic chains are usually added to transmit the torques from motors located on the base to the platform, in order to actuate a wrist. Packaging applications usually require five degrees of freedom, and two UPU chains are then used to actuate two rotational degrees-of-freedom (DOFs) on the end-effector. However, the UPU chain induces significant limitations for industrial use: it significantly constrains the workspace along the vertical direction and implies a backlash in the universal joints degrading the accuracy of the robot. In this paper, we investigate an alternative to the UPU kinematic chain for designing Delta-like robots with five DOFs. Indeed, the actuation of a two-DOFs wrist is performed through the use of a kinematic chain based on a succession of parallelograms associated with a Delta-like leg. After a description of the kinematic models of the modified leg and an analysis of its singularities, a design optimization procedure is presented in order to define suitable geometric parameters for a given industrial application. Finally, a prototype is presented and its performances are evaluated.
{"title":"Modeling and Design of a five Degrees-of-Freedom Delta-Like Robot for Fast Pick-and-Place Applications","authors":"Valentin Le Mesle, Vincent Bégoc, S. Briot","doi":"10.1115/1.4063359","DOIUrl":"https://doi.org/10.1115/1.4063359","url":null,"abstract":"\u0000 Delta-like architectures are widely used for fast pick-and-place applications. When rotational degrees of freedom are required to perform a task, one or more UPU kinematic chains are usually added to transmit the torques from motors located on the base to the platform, in order to actuate a wrist. Packaging applications usually require five degrees of freedom, and two UPU chains are then used to actuate two rotational degrees-of-freedom (DOFs) on the end-effector. However, the UPU chain induces significant limitations for industrial use: it significantly constrains the workspace along the vertical direction and implies a backlash in the universal joints degrading the accuracy of the robot. In this paper, we investigate an alternative to the UPU kinematic chain for designing Delta-like robots with five DOFs. Indeed, the actuation of a two-DOFs wrist is performed through the use of a kinematic chain based on a succession of parallelograms associated with a Delta-like leg. After a description of the kinematic models of the modified leg and an analysis of its singularities, a design optimization procedure is presented in order to define suitable geometric parameters for a given industrial application. Finally, a prototype is presented and its performances are evaluated.","PeriodicalId":50137,"journal":{"name":"Journal of Mechanical Design","volume":"151 1","pages":""},"PeriodicalIF":3.3,"publicationDate":"2023-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"73460428","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Power-split hybrid transmissions are the core components of power-split hybrid electric vehicles (HEV), and the quest for a more energy-efficient and higher-performing power-split hybrid transmission has long been the focus of study. In contrast to previously published methodologies, this paper proposes a novel approach for directly synthesizing power-split hybrid transmissions that makes use of the results of previously synthesized planetary gear trains (PGTs) rather than necessitating a re-synthesis of their PGTs. A new topological graph that can construct a bridge between the PGTs and power-split hybrid transmission has been developed, reducing the computational complexity of the synthesis process. The new topological graph is obtained by adding the topological characteristics of the power-split hybrid transmission to the PGT graph. A standard structure matrix is proposed to further screen out all the isomorphic configurations. The present method can generate various types of multi-PGT hybrid transmissions while avoiding mechanical and structural interference. The design process of configurations for power-split hybrid transmission with 3-column PGTs (3-PGT) is used as an example to prove the rationality of the method.
{"title":"Topological Graph Representation and Configuration Synthesis for Power split Hybrid Transmissions of Multi-Planetary Gear Trains","authors":"Meijie Geng, H. Ding, Tao Ke, Wenjian Yang","doi":"10.1115/1.4063287","DOIUrl":"https://doi.org/10.1115/1.4063287","url":null,"abstract":"\u0000 Power-split hybrid transmissions are the core components of power-split hybrid electric vehicles (HEV), and the quest for a more energy-efficient and higher-performing power-split hybrid transmission has long been the focus of study. In contrast to previously published methodologies, this paper proposes a novel approach for directly synthesizing power-split hybrid transmissions that makes use of the results of previously synthesized planetary gear trains (PGTs) rather than necessitating a re-synthesis of their PGTs. A new topological graph that can construct a bridge between the PGTs and power-split hybrid transmission has been developed, reducing the computational complexity of the synthesis process. The new topological graph is obtained by adding the topological characteristics of the power-split hybrid transmission to the PGT graph. A standard structure matrix is proposed to further screen out all the isomorphic configurations. The present method can generate various types of multi-PGT hybrid transmissions while avoiding mechanical and structural interference. The design process of configurations for power-split hybrid transmission with 3-column PGTs (3-PGT) is used as an example to prove the rationality of the method.","PeriodicalId":50137,"journal":{"name":"Journal of Mechanical Design","volume":"107 1","pages":""},"PeriodicalIF":3.3,"publicationDate":"2023-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"79294093","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Recently, online user-generated data has emerged as a valuable source for consumer product research. However, most studies have neglected the brand effect, although it is a significant factor in conventional market research. This paper demonstrates the importance of brands in data-driven design using online reviews. Specifically, this study utilizes game theory and suggests a game setting representing market competition. Elements of the game are determined based on online data analysis. The proposed approach consists of three stages. The first stage divides online customers into different segments and analyzes them to extract the feature importance of each brand in each segment. The importance is based on the positive term frequency of features, and it becomes the customer’s partial utility for each feature. The second stage defines the specification of product candidates and calculates their costs. This study refers to real market datasets (Bill of Materials) available online. At this point, the game is all set. The final stage finds the Nash Equilibrium of the designed game and compares the optimal strategy for a product portfolio with and without brand consideration. The suggested approach was tested on smartphone reviews from Amazon. The result shows that the lack of brand consideration leads a company to choose a non-optimal product strategy, illustrating the significance of the brand factor. Keywords: data-driven design, online review, brand effect
{"title":"Analysis of Brand Effects in Data-Driven Design Based on Online Reviews","authors":"Seyoung Park, Harrison M. Kim","doi":"10.1115/1.4063288","DOIUrl":"https://doi.org/10.1115/1.4063288","url":null,"abstract":"\u0000 Recently, online user-generated data has emerged as a valuable source for consumer product research. However, most studies have neglected the brand effect, although it is a significant factor in conventional market research. This paper demonstrates the importance of brands in data-driven design using online reviews. Specifically, this study utilizes game theory and suggests a game setting representing market competition. Elements of the game are determined based on online data analysis. The proposed approach consists of three stages. The first stage divides online customers into different segments and analyzes them to extract the feature importance of each brand in each segment. The importance is based on the positive term frequency of features, and it becomes the customer’s partial utility for each feature. The second stage defines the specification of product candidates and calculates their costs. This study refers to real market datasets (Bill of Materials) available online. At this point, the game is all set. The final stage finds the Nash Equilibrium of the designed game and compares the optimal strategy for a product portfolio with and without brand consideration. The suggested approach was tested on smartphone reviews from Amazon. The result shows that the lack of brand consideration leads a company to choose a non-optimal product strategy, illustrating the significance of the brand factor. Keywords: data-driven design, online review, brand effect","PeriodicalId":50137,"journal":{"name":"Journal of Mechanical Design","volume":"90 1","pages":""},"PeriodicalIF":3.3,"publicationDate":"2023-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"76337955","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Christopher S. Mabey, Erin Peiffer, Nordica A. MacCarty, Christopher A. Mattson
This paper presents a methodology for predicting the adoption and social impact of a product using agent-based modeling (ABM) and neural networks to aid in decision-making related to the design and implementation of the product in a sociotechnical system. The collection of primary data on the social impact of a product is also outlined. Although this paper illustrates the method for improved cookstoves in Uganda, the method can be applied to a wide range of contexts. A field study was carried out in Uganda, consisting of two phases of data collection. The data from the fieldwork was used to train a neural network to predict if an individual would adopt an improved cookstove. Data collected from surveys and the trained adoption model were used to create an ABM to estimate adoption rates and social impacts experienced by households that had adopted technology and to assess social impact indicators. The contributions of this article are a method for collecting primary social impact data on a product and how to integrate those data into a predictive agent-based social impact model. This methodology also enables the examination of leverage points in the sociotechnical system to improve the social impact of a product as it is implemented in society.
{"title":"Simulating the Adoption and Social Impact of Improved Cookstoves in Uganda Using Agent-Based Modeling and Neural Networks","authors":"Christopher S. Mabey, Erin Peiffer, Nordica A. MacCarty, Christopher A. Mattson","doi":"10.1115/1.4063237","DOIUrl":"https://doi.org/10.1115/1.4063237","url":null,"abstract":"\u0000 This paper presents a methodology for predicting the adoption and social impact of a product using agent-based modeling (ABM) and neural networks to aid in decision-making related to the design and implementation of the product in a sociotechnical system. The collection of primary data on the social impact of a product is also outlined. Although this paper illustrates the method for improved cookstoves in Uganda, the method can be applied to a wide range of contexts. A field study was carried out in Uganda, consisting of two phases of data collection. The data from the fieldwork was used to train a neural network to predict if an individual would adopt an improved cookstove. Data collected from surveys and the trained adoption model were used to create an ABM to estimate adoption rates and social impacts experienced by households that had adopted technology and to assess social impact indicators. The contributions of this article are a method for collecting primary social impact data on a product and how to integrate those data into a predictive agent-based social impact model. This methodology also enables the examination of leverage points in the sociotechnical system to improve the social impact of a product as it is implemented in society.","PeriodicalId":50137,"journal":{"name":"Journal of Mechanical Design","volume":"45 1","pages":""},"PeriodicalIF":3.3,"publicationDate":"2023-08-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"73810242","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Design heuristics are traditionally used as qualitative principles to guide the design process, but they have also been used to improve the efficiency of design optimization. Using design heuristics as soft constraints or search operators has been shown for some problems to reduce the number of function evaluations needed to achieve a certain level of convergence. However, in other cases, enforcing heuristics can reduce diversity and slow down convergence. This paper studies the question of when and how a given set of design heuristics represented in different forms (soft constraints, repair operators, biased sampling) can be utilized in an automated way to improve efficiency for a given design problem. An approach is presented for identifying promising heuristics for a given problem by estimating the overall impact of a heuristic based on an exploratory screening study. Two impact indices are formulated: weighted influence index and hypervolume difference index. Using this approach, the promising heuristics for 4 design problems are identified and the efficacy of selectively enforcing only these promising heuristics over both enforcement of all available heuristics and not enforcing any heuristics is benchmarked. In all problems, it is found that enforcing only the promising heuristics as repair operators enables finding good designs faster than by enforcing all available heuristics or not enforcing any heuristics. Enforcing heuristics as soft constraints or biased sampling functions results in improvements in efficiency for some of the problems. Based on these results, guidelines for designers to leverage heuristics effectively in design optimization are presented.
{"title":"Identifying and Leveraging Promising Design Heuristics for Multiobjective Combinatorial Design Optimization","authors":"Roshan Suresh Kumar, Emilie Baker, Srikar Srivatsa, Meredith Silberstein, Daniel Selva","doi":"10.1115/1.4063238","DOIUrl":"https://doi.org/10.1115/1.4063238","url":null,"abstract":"\u0000 Design heuristics are traditionally used as qualitative principles to guide the design process, but they have also been used to improve the efficiency of design optimization. Using design heuristics as soft constraints or search operators has been shown for some problems to reduce the number of function evaluations needed to achieve a certain level of convergence. However, in other cases, enforcing heuristics can reduce diversity and slow down convergence. This paper studies the question of when and how a given set of design heuristics represented in different forms (soft constraints, repair operators, biased sampling) can be utilized in an automated way to improve efficiency for a given design problem. An approach is presented for identifying promising heuristics for a given problem by estimating the overall impact of a heuristic based on an exploratory screening study. Two impact indices are formulated: weighted influence index and hypervolume difference index. Using this approach, the promising heuristics for 4 design problems are identified and the efficacy of selectively enforcing only these promising heuristics over both enforcement of all available heuristics and not enforcing any heuristics is benchmarked. In all problems, it is found that enforcing only the promising heuristics as repair operators enables finding good designs faster than by enforcing all available heuristics or not enforcing any heuristics. Enforcing heuristics as soft constraints or biased sampling functions results in improvements in efficiency for some of the problems. Based on these results, guidelines for designers to leverage heuristics effectively in design optimization are presented.","PeriodicalId":50137,"journal":{"name":"Journal of Mechanical Design","volume":"54 1","pages":""},"PeriodicalIF":3.3,"publicationDate":"2023-08-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"90901864","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This study proposes an inverse method for synthesizing shape-morphing structures in the lateral direction by integrating two-dimensional hexagonal unit-cell with curved beams. Analytical expressions are derived to formulate the effective Young's modulus and Poisson's ratio for the base unit cell as a function of its geometric parameters. The effective lateral Poisson's ratio can be controlled by manipulating a set of geometric parameters, resulting in a dataset of over 6000 data points with Poisson's ratio values ranging from -1.2 to 10.4. Furthermore, we utilize the established dataset to train an inverse design framework that utilizes a physics-guided neural network algorithm, and the framework can predict design parameters for a targeted shape-morphing structure. The proposed approach enables the generation of structures with tailored Poisson's ratio ranging from -1.2 to 3.4 while ensuring flexibility and reduced stress concentration within the predicted structure. The generated shape-morphing structures' performance is validated through numerical simulation and physical tensile testing. The FEA simulation results confirm agreement with the designed values for the shape-morphing structure, and the tensile testing results reveal the same trend in shape-morphing behavior. The proposed design automation framework demonstrates the feasibility of creating intricate and practical shape-morphing structures with high accuracy and computational efficiency.
{"title":"Inverse Design of 2D Shape-Morphing Structures","authors":"Mohammad Abu-Mualla, Victor Jiron, Jida Huang","doi":"10.1115/1.4063191","DOIUrl":"https://doi.org/10.1115/1.4063191","url":null,"abstract":"\u0000 This study proposes an inverse method for synthesizing shape-morphing structures in the lateral direction by integrating two-dimensional hexagonal unit-cell with curved beams. Analytical expressions are derived to formulate the effective Young's modulus and Poisson's ratio for the base unit cell as a function of its geometric parameters. The effective lateral Poisson's ratio can be controlled by manipulating a set of geometric parameters, resulting in a dataset of over 6000 data points with Poisson's ratio values ranging from -1.2 to 10.4. Furthermore, we utilize the established dataset to train an inverse design framework that utilizes a physics-guided neural network algorithm, and the framework can predict design parameters for a targeted shape-morphing structure. The proposed approach enables the generation of structures with tailored Poisson's ratio ranging from -1.2 to 3.4 while ensuring flexibility and reduced stress concentration within the predicted structure. The generated shape-morphing structures' performance is validated through numerical simulation and physical tensile testing. The FEA simulation results confirm agreement with the designed values for the shape-morphing structure, and the tensile testing results reveal the same trend in shape-morphing behavior. The proposed design automation framework demonstrates the feasibility of creating intricate and practical shape-morphing structures with high accuracy and computational efficiency.","PeriodicalId":50137,"journal":{"name":"Journal of Mechanical Design","volume":"80 1","pages":""},"PeriodicalIF":3.3,"publicationDate":"2023-08-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"74433601","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：